How to Create Subtitles for a YouTube Video Without Captions

When a video has no caption track, extraction tools have nothing to download. The solution is transcription: create text from authorized audio, segment it into cues, synchronize those cues, and review the result. Automatic speech recognition can accelerate the first draft, but names, numbers, punctuation, and timing still need human verification.

Prepare the media and reference material

Work from the final video cut whenever possible. A subtitle track made against a rough cut will drift or jump after edits. Collect the script, slide deck, speaker names, product names, URLs, and an approved glossary. Note whether the deliverable is dialogue subtitles or accessibility captions that also require sound effects and speaker labels.

Use clean audio from a video you own or are authorized to process. If several speakers overlap, a mixed track may be insufficient; request isolated channels when available.

Create the transcript draft

Choose manual transcription, speech recognition, or a hybrid based on audio quality and risk. For technical, medical, legal, or safety-critical content, do not accept recognized numbers and terminology without source verification.

Draft transcript:

00:00:08 speaker 1 today we'll compare s r t and web v t t
00:00:13 speaker 1 keep the original before converting

Correct text before fine timing:

Today we'll compare SRT and WebVTT.
Keep the original before converting.

Segment into cues

Break at sentence or clause boundaries and align with speech. Avoid showing the next idea before it is spoken. Keep line length, duration, and reading speed within the delivery specification.

1
00:00:08,200 --> 00:00:12,600
Today we'll compare SRT and WebVTT.

2
00:00:13,050 --> 00:00:15,900
Keep the original before converting.

For dense speech, concise editing may be necessary. Preserve meaning and required terminology; do not delete qualifiers simply to fit.

Synchronize in two passes

In the rough pass, place cue starts and ends around the correct utterances. In the fine pass, check lead-in, end lag, shot changes, overlaps, and reading time. Measure the opening, middle, and ending for global offset or drift. Use the timing synchronization guide if errors follow a pattern.

Add caption information when required

Accessibility captions may need [applause], [phone rings], lyrics, and speaker identification. Describe sounds that convey meaning, not every incidental noise. Use consistent brackets and speaker style.

Export and validate

Export UTF-8 SRT for broad compatibility or WebVTT for a web player. Review SRT vs VTT before choosing. Upload an unlisted test, enable captions, and watch on desktop and mobile. Verify the first cue, names and numbers, rapid sections, scene changes, and the last cue.

Run the subtitle quality checklist before publication. Once a source-language track is approved, the complete workflow continues through translation and optional TTS or dubbing preparation.